基于小波变换的脉象信号特征提取方法

为了较好地区分正常人与心脏病人的脉象信号，利用小波变换奇异性检测功能与多尺度分辨特性，提出了两种提取脉象信号特征的方法：连续小波变换法和二进小波变换法。在此基础上，构造了两种特征向量：小波变换系数的尺度——主波峰值和小波变换的尺度——能量值。经过对临床采集的235例脉象信号的处理与分析统计，所得数据具有较好的重复性与稳定性，可以作为用于脉象信号识别的特征向量。

Feature Extraction Methods for Pulse Signal Based on Wavelet Transform

In order to distinguish the pulse signals of the normal from the disease ones, two methods for extracting features are proposed based on the singularity theory and the multi-scale resolution character of the wavelet transform. They are continuous wavelet transform method and discrete wavelet transform method. According to the two methods, the scale-wave crest amplitude and the scale-energy are constructed. By processing and analyzing 235 pulse signals, the statistics show that the data have better repeatability and stability. So they can be regarded as the feature vectors to identify the pulse signals.